CCM looks at how card issuers can effectively use invitation-to-apply marketing campaigns to acquire new customers without exposing their portfolios to unwanted risk.
Invitation-to-apply marketing campaigns are becoming increasingly important as credit card issuers seek new opportunities to augment traditional direct mail.
Traditional campaigns have been largely dependent on preapproved credit offers, which are now starting to experience response-rate compression. According to a report last June from Chicago-based Synovate's Mail Monitor research service, direct-mail volumes for credit card issuers over the last decade increased 4.7 times-915 million solicitations in 1992 to 4.3 billion pieces in 2003-while total responders changed less than 1%-25.64 million in 1992 to 25.75 million in 2003-during the same period. The vast majority of these offers were preapproved.
The growing appeal of invitation-to-apply campaigns revolves around the underlying opportunity to reach new markets. Invitation-to-apply campaigns are those that do not rely on credit data for marketing, so the campaigns can invite consumers to apply for a card without making the firm offers of credit that a preapproved offer must have.
Most Americans are more familiar with preapproved offers because they receive several in the mail each month. Preapproved offers depend on credit-bureau data and are required by the federal Fair Credit Reporting Act (FCRA) to contain a guaranteed offer of credit. The only escape clause is if the issuer discovers that the respondent to a preapproved offer experienced a drastic decline in creditworthiness, such as bankruptcy, between the time the list of prescreened potential customers was compiled and the prospect's response.
This article looks at a new market opportunity that is of particular interest to credit card marketers: the potential to effectively market to a prospect universe of an estimated 54 million untapped and underserved consumers. This market segment includes many consumers who are credit eligible yet are missed by traditional preapproved offer campaigns. (The U.S. Census Bureau estimates the country has 215 million credit-eligible individuals; however, Fair Isaac research estimates that only 74.8% meet preapproved credit standards and 25.2%, or 54 million, are eligible but are missed in preapproved campaigns because traditional credit data are not available.)
First we will explore how invitation-to-apply marketing campaigns have historically suffered from adverse selection when targeting the underserved market because the lowest performing consumers are the most likely to respond. Next we will discuss policies and strategies that card issuers can use to improve their campaigns if they factor in likelihood of approval.
We also will look at an example that illustrates the opportunities for combined savings and increased revenues across marketing campaigns for card issuers if a rigorous analytic strategy is adopted for invitation-to-apply campaigns. We will conclude by summarizing how new, enhanced analytics ultimately will be advantageous to consumers as card issuers look to adopt the same analytic strategies into their underwriting practices and policies.
One of the most common direct-marketing tactics employed by the card industry is the use of credit scores to generate preapproved marketing offers. A relatively simple segmentation of consumers' credit scores allows marketers to tier targeted lists for a wide variety of credit product promotions. Because of the proven analytic models that generate credit scores, marketers have become confident enough of their predictive abilities to create broad, preapproved campaigns that make guaranteed (according to FCRA standards) offers of credit to individuals who pass their prescreening criteria.
While this is a good means for segmentation, we estimate, as noted above, that there are 54 million people who do not have sufficient credit history to produce a credit score and are being missed by preapproved offers. This results in a classic Catch-22: consumers who already have credit are receiving many additional offers for cards that are unneeded or unwanted, while consumers with the most desire and need for credit access-because they don't yet have it-are not receiving any offers.
There is great diversity among the estimated 54 million "underserved" people. They range from wealthy or independent individuals who always pay cash for goods and services, to recently divorced adults whose credit was in their spouses' name, to highly educated immigrants whose credit history from their home country does not transfer into the U.S. credit system.
While many of these underserved consumers would be attractive prospects for lenders, preapproved credit card marketing campaigns miss them because the consumers don't yet have a credit history or credit score. At the other extreme, some of the 54 million prospects will pose high credit risk to lenders so issuers would want to identify them for separate treatment.
Tremendous efficiency would be gained if lenders could look at their prospect list and identify those who would be likely to respond to offers and who would be approved in a subsequent credit decision. For example, a lender eliminating just 30% of its mailings (which cost about 45 cents per prospect) to 10 million prospects could save more than $1.3 million on the campaign by systematically eliminating the prospects that are least likely to respond and least likely to be approved when the credit decision is made. (Here's the math: 10 million prospects times 30% times 45 cents equals $1.35 million.)
Lenders should model the kind of prospects they want to market to and use available data to help locate those prospects. Rather than simply relying on high-level "market clusters," lenders should apply advanced analytic modeling to information on their prospect universe that has been carefully selected.
To identify their ideal prospects from outside sources-such as the well-known demographic compilers as well as sources of public information such as tax liens, property information and geographic data-card issuers can apply the same analytic models to their current customer portfolio in order to find correlations between customer behaviors and different demographic details. Then, by taking into account similar information from outside sources on known accounts, a card issuer can make predictions about how prospects will behave.
To produce useful results, the sample of accounts needs to be sufficiently large and representative of all demographic types. This is a challenge not only for mid-sized card issuers but also for many of the nation's largest. To address this challenge, many issuers provide information to data consortiums to create pooled data models.
This work can provide enormous value by improving the ability of card issuers to suppress mailings to those who are unlikely to respond and those who would be eliminated during underwriting. As mentioned earlier, hard costs are saved during the initial marketing phase-45 cents per mailing in our example. Also, the resources used during underwriting, approximately $9 per account, are more effectively assigned to desired accounts.
In the chart on page 27, we see actual figures about the amount of savings that card issuers can realize as they begin to properly allocate marketing dollars with better prospect segmentation. This case reviews how using an analytic score-such as one designed to use non-traditional data-can effectively reduce mailings and underwriting costs while improving the booking rate and quality of accounts, as well as increasing the net lifetime value of those accounts. The numbers cited here are based on the results of one company's campaign using traditional segmentation tools, and project the value that an analytic score, in this case the Qualify score, would have provided had it been used in addition to the traditional tools.
Instead of mailing to all 10 million prospects, the segmentation called for mailing to only the top 50% of the prospects. This immediately reduced mailing costs by $2.2 million while maintaining a similar overall response rate, with a higher number of accounts booked and a reduction in total underwriting costs.
While it eliminated some of the higher risk during the modeling phase, the approach resulted in an overall 15% improvement in the quality of the booked accounts as measured by delinquencies. Perhaps most impressive, once all of the factors were combined and the total net lifetime value of the portfolio was calculated, the card issuer increased the new accounts' net lifetime value by 40%. Net lifetime value means the net present value of the customer based on projections of cash flow over the duration of the customer's account, assuming a five-year retention rate.
While this article has addressed the opportunity for card issuers to improve their marketing effectiveness, lenders must develop a process to underwrite credit offers to consumers who respond to invitation-to-apply offers. Using many of the same data sources and strategies, lenders are developing enhanced analytic applications that will allow them to confidently lend to credit applicants who do not appear in the most commonly used credit bureaus.
As issuers roll out new underwriting solutions, underserved consumers will begin to benefit from the opportunity to establish credit and receive better rates from lenders for two primary reasons. First, as lenders gain experience and confidence in these new scoring models and methodologies, they will properly fine-tune their risk premiums.
Moreover, as more lenders become familiar with such advancements in scoring, a more competitive environment could emerge in which more credit card issuers vie for the same consumers. The end result for the consumer will be increased borrowing choices, more attractive financing options and overall better interest rates.
Craig Dillon is responsible for Minneapolis-based Fair Isaac Corp.'s Global Scoring Solution product offerings, with a focus on applying analytics and advanced technology to both existing and new business opportunities. Before joining Fair Isaac in 2001, he served as chief technology and operating officer at ieWild Inc., where he oversaw all product development and operations. Earlier, he was Advanced Projects and Business Operations Director at HNC Software Inc. He holds a Ph.D in computer science from Curtin University, Australia, and bachelor's degrees in electrical engineering and science from Melbourne University, Australia.
Authoritative analysis and perspective for every segment of the payments industry
Authoritative analysis and perspective for every segment of the industry
Have an account? Sign In